paperKB
coga / coga-kb
Processing
Help
Sign in

Chunk #19 — 3. RESULTS — 3.1 Descriptive Statistics — 3.1.1 Measurement Invariance

Source
Is the Fagerström test for nicotine dependence invariant across secular trends in smoking? A question for cross-birth cohort analysis of nicotine dependence.
Embedded
yes

Text

Evaluating item-level non-invariance in the model using birth cohort <1945 as the reference, 10 parameters show statistically significant measurement non-invariance. Time to first cigarette in the morning was the only current FTND item that did not show measurement non-invariance. The effect sizes for these direct paths ranged from r2=0.0003 to r2=0.0035 indicating that, despite statistical significance, little variance (<0.5%) of the model was explained by measurement non-invariance. Fewer significant parameters were noted for the model using birth cohort 1945–1975 as the reference group, and they were of lower magnitude (Table S2)1. A similar pattern of results was observed for lifetime FTND.